JP2017530425A - Method and system for supporting improvement of user utterance in a specified language - Google Patents

Abstract

A system and method for assisting in improving a user's utterance in a specified language, the method receiving from the user text in the specified language or other language for the user to utter in the specified language; Processing the text to derive one or more expected characteristics of the text utterance in the specified language; and receiving a voice signal of the user speaking the text in the specified language Processing the speech signal to derive one or more characteristics of the user speaking the text; and determining the accuracy of the user speaking the text; The one or more characteristics of the user speaking the text and a corresponding characteristic of the one or more expected characteristics of the utterance of the text; Comprising the steps of: compare, and outputting the feedback indicating the accuracy to the user, the.

Description

  The present invention relates to a method and system for helping to improve a user's utterance in a specified language, and more particularly to receiving from a user text in a specified language or other language for the user to utter in the specified language.

  The invention relates to the use of comparing the characteristics of a user who is speaking text with the corresponding expected characteristics of the text utterance to determine the accuracy of the user speaking the text, and their accuracy Is used to output feedback indicating to the user, and these uses are unique but do not exclude other uses.

  Traditionally, those who wish to learn a desired language have taken a suitable curriculum with teachers and other students. In this educational process, for example, the teacher provides the student with text to speak and provides feedback to the student on how the student spoke the text compared to the expected utterance of the text. . This feedback may include reviews regarding specific characteristics of the text utterance, such as fluency and pronunciation. In addition, teachers can provide assignments for students to practice at home. However, in that case, the student does not receive any feedback at home, and in some situations, fosters bad habits that impair the student's ability to learn and improve utterances in the desired language.

  In existing examples, an electronic utterance feedback system is employed to eliminate the need for physical teachers for students to learn utterances in a desired language. In this existing example, each word having known utterance characteristics such as pronunciation is presented to the user for the user to read it aloud as a word. The feedback system receives the speech signal of a user speaking each predetermined word and applies a speech recognition algorithm to determine whether the user has spoken those words correctly. However, in this example, the user may feel that each given and often repeated word is boring and meaningless, and thus may lose interest in improving the user's utterance in the desired language. There is.

  According to a first aspect of the present invention, a method is provided for assisting in improving a user's utterance in a specified language, the method comprising text in a specified language or other language for the user to speak in the specified language. Receiving from the user; processing the text to derive one or more expected characteristics of the speech of the specified language text; and receiving a voice signal of the user speaking the text of the specified language. Processing the speech signal to derive one or more characteristics of the user speaking the text, and speaking the text to determine the accuracy of the user speaking the text Comparing the one or more characteristics of the user who is playing with a corresponding one of the one or more expected characteristics of the utterance of text, and a feedback indicating the accuracy. Comprising a step of outputting a click to the user, the.

  In one embodiment, the one or more expected characteristics of the text utterance and the one or more characteristics of the user speaking the text include at least one of pronunciation, fluency, and prosodic features. Including. Prosodic features of utterance include changes in syllable length, voice loudness, and voice pitch. One skilled in the art will appreciate that expected pronunciation, fluency, and prosodic features can be determined by empirical analysis of the speaker's utterances in the specified language. Certainly, prosodic features can indicate the accent of a specified language. For example, in this case, as an example, the prosodic features of the American Midwestern American accent are determined by empirical analysis, and this method determines feedback indicating the accuracy of the user speaking the Midwestern American accent. It can be output to the user.

  One skilled in the art will recognize that the accuracy determination for each of the characteristics of the user speaking the text is determined by the determination of the accuracy of the user speaking the text, along with a corresponding one of the expected characteristics of the text speaking. It will be appreciated that it includes determining a measure of the quality of the user speaking the text based on the sum (eg, 1/100 to 100/100). Thus, in this embodiment, the method determines the accuracy of each specified characteristic, such as the speed, voice pitch, energy, pronunciation, and fluency of the user speaking the text. Determine the accuracy of the user who is speaking.

  In one embodiment, the method further comprises parsing the text into one or more sentences or clauses of words. For example, a collection of words can be parsed into one or more sentences by sequentially searching the text containing those words for the appearance of ending citations such as terminators, question marks, and exclamation marks. Do.

  As an alternative, the word aggregation performs a syntactic analysis on a set of fixed word number clauses that are typically two to eight words in size. Subsequently, each phrase is output and taught to the user in turn, but one phrase is taught first, then a second phrase is added, and taught in this way until the entire collection of words is taught. In one variation, phrases can be constructed from left to right from the beginning of the word aggregate until the entire aggregate is taught. As an alternative, a clause can be constructed from the end point of the word aggregate to its presence from right to left, eg, if the number of words in the clause is three, first teach the last three words of the aggregate. Then teach the last 6 words, then the last 9 words, and so on until the aggregation is complete.

  In other embodiments, a grammar for one or more sentences of the text is also generated. One skilled in the art will appreciate that a grammar (eg, grammars) is a speech recognition grammar, ie, a formal grammar structure that can be recognized by a speech recognition device that executes a speech recognition algorithm. In one example, a reminder signal for letting the user speak the first sentence of each sentence is provided to the user via the text of the first sentence of each sentence displayed on the display of the user device. To do. The method receives an audio signal of a user speaking the first sentence of each sentence of the text and subsequently indicates the accuracy of the user speaking the first sentence of each sentence. The method further includes outputting feedback to the user. After outputting the feedback, the user is prompted to speak the second sentence of each sentence.

  Thus, in one example of use, the user first enters text in a specified language or other language that he wishes to master speech, and the text in the other language is translated into the specified language for display to the user. The text is parsed into each sentence.

  A user who wishes to improve speech in the specified language is prompted to speak the first sentence with the text displayed. The voice signal of the user who is speaking the sentence is received and processed by the voice recognition device so that the user can receive feedback related to the pronunciation and fluency of the sentence. After receiving the feedback, the user can start the next sentence, can receive the feedback as well, and so on.

  For example, the method uses a speech recognition grammar specification (SRGS) to control all recognition operations with a speech recognition device that executes a speech recognition algorithm. SRGS is the World Wide Web Consortium (W3C) standard for speech recognition grammar. It will be appreciated that the speech recognition grammar is a set of word patterns that give instructions to the speech recognizer as to what a human expects to say.

  In one embodiment, the method further includes processing the text to make the text a specified language voice and outputting the voice to the user via a speaker of the user's device. That is, in this embodiment, the text for the speech generator is employed to assist the user in improving the utterance of the specified language. For example, in use, the user speaks a sentence, receives feedback regarding the accuracy of how the user speaks the sentence, and listens to the expected utterance of the sentence for comparison. Certainly, the order in which the method is used is, for example, (a) input text, input utterance, provide output of expected utterance, or (b) input text, expected It is possible to provide an utterance output and provide an utterance input for comparison.

  One skilled in the art will recognize that the step of processing the speech signal to derive one or more characteristics of the user speaking the text is performed using a speech recognition algorithm and one or more expected text utterances. It will be appreciated that the step of processing the text to derive the characteristics is performed using a speech synthesis algorithm.

  The accuracy preferably includes two or more confidence values associated with the accuracy of one or more characteristics of the user speaking the text corresponding to one or more expected characteristics of the text utterance. The confidence value can be, for example, an array of “pass” / “fail” based on how close the user-spoken characteristic is to the expected characteristic. For example, we set a baseline of 70% match between the fluency and pronunciation of speech and the expected fluency and pronunciation of others speaking the text, and the “pass” value is the spoken value of the text Determined by whether fluency and pronunciation exceeded 70% agreement. In one example, the confidence value associated with accuracy is “very accurate (eg, greater than 80% match)”, “slightly accurate (eg, 50% to 80% match)”, “slightly inaccurate (eg, 20% % To 50% match) "and" very inaccurate (eg 0% to 20% match) ". Referring to this example, the feedback also includes a color that indicates a variety of accuracy confidence values. For example, green for "very accurate", orange for "slightly accurate", brown for "slightly incorrect", and red for "very inaccurate" On the display. It is believed that audio feedback can also be provided to the user, such as outputting audio “very inaccurate” through the user device's speaker. Other forms of feedback include numerical grades (eg, 1 to 10), written grades (eg, A to F), badges, or some other visual indicator of feedback.

  In other embodiments, the method further includes processing the text to derive the prohibited word so that the predetermined prohibited word in the text can be censored. For example, spoken words, words that indicate hate speech, etc. are pre-determined and accessible in memory before the steps of parsing text into sentences are performed.

  In one embodiment, the method further includes recording in a memory the voice of the user speaking the text in the specified language. In this way, the method can process, for example, examples of speech recorded in a database to derive one or more expected characteristics of a text utterance. That is, for example, the expected pronunciation of a text can be determined by analyzing the recordings of many users speaking the text. In addition, the recorded utterances of a particular user can be used for later analysis, such as showing the progress of that user speaking a specified language. In a further example, all text entered and received by the user is retained for offline analysis. That is, search text string collections for iterations (eg, the same input from different users) and for specific words or themes that many users are familiar with.

  In accordance with another aspect of the invention, a system is provided for assisting in improving a user's utterance in a specified language, the system being a processor, the specified language for the user speaking in the specified language, or An input module configured to receive text in another language from the user and receive the voice of the user speaking the text in the specified language, and one or more expected utterances of the text in the specified language Process the text to derive characteristics, process the speech to derive one or more characteristics of the user speaking the text, and determine the accuracy of the user speaking the text Therefore, a process configured to compare the one or more characteristics of a user uttering text with a corresponding one of the one or more expected characteristics of the utterance of text Including a Joule, a processor and an output module configured to output to a user feedback indicating accuracy.

  In accordance with another aspect of the present invention, a system is provided for assisting in improving a user's speech in a specified language, the system configured to display text in the specified language or other languages. A text input means configured to input the text in the specified language or another language for the user to speak in the specified language, and to input the voice of the user speaking the text in the specified language A microphone configured and a processor, the input module configured to receive text input in a specified language or other language, and to receive a voice of a user speaking the text in the specified language; One or more of the users who are processing the text to derive one or more expected characteristics of the speech of the text in the specified language and that are speaking the text The one or more characteristics of the user speaking the text are determined for the utterance of the text to process the speech to derive the characteristics of the text and to determine the accuracy of the user speaking the text A processor having a processing module configured to compare with a corresponding one of the one or more expected characteristics and an output module configured to output feedback indicating accuracy to a user; ,including.

  In one embodiment, the system further includes a server including the processor in data communication over a network with a user device including the display, the text input means, and the microphone. In other words, in this embodiment, a user can use a user device (for example, a tablet, a personal computer, Or a smartphone). Thus, the user enters text to speak in a specified language, for example by typing text or by electronically pasting text from a document into an assigned text box. The text is sent over the network to the server, which parses the text into each sentence and outputs a first sentence for the user to speak over the network for display on the user device display. . The user can speak the first sentence, and the microphone of the speaking user's voice is captured and sent to the processor for processing. Thereby, as described above, the feedback can be determined, and can be output to the user via the display of the user device, for example. That is, the display of the user device is further configured to display feedback to the user.

  In other embodiments, a user device that includes a display, text input means, and a microphone also includes the processor. Thus, in this embodiment, the processing is performed locally on the user device.

  Preferably, the processing module is further configured to parse the text into one or more sentences, and the output module is further configured to output a reminder signal to cause the user to speak the first sentence of each sentence. . Similarly, the input module is further configured to receive the voice of a user speaking the first sentence of each sentence of the text, and the output module is one or more of the utterances of the first sentence of each sentence of the text. And further configured to output feedback to the user indicating the accuracy of one or more characteristics of the user speaking the first sentence of each sentence of the text corresponding to the expected characteristics of the user.

  As described in some embodiments, the system allows user-generated text to be used in assisting the user in improving speech in a specified language. The text is parsed into each sentence, and the user is prompted to speak one sentence at a time so that feedback can be received and reviewed by the user upon completion of each sentence. Thus, in this case, the output module outputs a feedback indicating the accuracy of one or more characteristics of the user speaking the first sentence of each sentence of the text to the user after the second of each sentence. It is further configured to output a further prompting signal for causing the user to speak the sentence.

  Similarly, the processing module is further configured to generate a grammar for one or more sentences of the text, and the reminder signal includes the text of the first sentence of each sentence.

  In one embodiment, the processing module is further configured to process the text to derive an expected utterance of the text in the specified language. In this embodiment, the output module is configured to output an expected utterance of text for output to the user via the speaker of the user device. For example, the processing module includes a speech synthesis module for processing text to derive an expected utterance to output to the user, for example after the user utters a sentence. This allows the user to compare user utterances and further improve them.

  In one embodiment, the processing module is further configured to process the text to derive a predetermined prohibited word in the text such that the prohibited word can be censored as described above. The forbidden words can be located away from the processor and stored in a memory that can be accessed over the network or locally located.

  According to another aspect of the invention, computer program code is provided that, when executed, performs the method described above.

  According to another aspect of the invention, a tangible computer readable medium is provided that includes the program code described above.

  According to another aspect of the present invention, a data file including the program code described above is provided.

In order that the present invention may be more clearly understood, examples of embodiments will be described with reference to the accompanying drawings.
FIG. 4 is a flowchart illustrating a method for supporting improvement of a user's utterance in a specified language according to an exemplary embodiment of the present invention. 1 is a diagram illustrating an outline of a system for supporting improvement of a user's utterance in a specified language according to an embodiment of the present invention; FIG. FIG. 3 shows a further overview of the system of FIG. 2 showing a system communicating with user devices over a network.

  In accordance with one embodiment of the present invention, a method 10 is provided that assists in improving user utterances in the specified language shown in FIG. The method 10 includes receiving 12 from a user text in a specified language or other language for the user to speak in the specified language, and deriving one or more expected characteristics of the speech of the text in the specified language. Processing the text, step 14 receiving the voice of the user speaking the text in the specified language, and processing the voice to derive one or more characteristics of the user speaking the text. Step 18 and corresponding one or more characteristics of the user speaking the text to one or more expected characteristics of the text speaking to determine the accuracy of the user speaking the text. Comparing to the characteristic 20 includes outputting a feedback indicating accuracy to the user 22.

  As described above, the one or more expected characteristics of the text utterance and the characteristics of the user speaking the text include at least one of pronunciation, fluency, and prosodic features. Prosodic features of utterances include changes in syllable length of words, loudness, and loudness. Thus, in an example of use, to determine the accuracy of the user speaking the text, the fluency of the expected utterance of the input text and, for example, the pronunciation is derived from the text and the text is spoken. Compared to the fluency and pronunciation derived from the voice received by the user. The accuracy of the speaking user regarding pronunciation and fluency is fed back to the user to assist the user in improving the user's utterance in the specified language (eg, English).

  In another embodiment of the present invention, the system 24 shown in FIG. 2 is provided to assist in improving the user's utterance in the specified language performing the method 10. System 24 includes a processor 26 having a number of modules for performing method 10. That is, the processor 26 includes an input module 28, a processing module 30, and an output module 32. The processor 26 may receive and transmit information to and from other components of the system 24, such as the memory 52 (shown in FIG. 3) over a network and / or over the communication channel 34, for example. Arrange for configuration. In this embodiment, processor 26 is executed by a computer that communicates via communication channel 34 with input and output devices included in user device 38 (shown in FIG. 3). However, as described above, it is conceivable that the memory 52 can be located in a server remote from the processor 26 accessible via the network, for example. In any event, those skilled in the art will recognize that each module of input 28 and output 32 has a suitable interface for connection to each module in the network and system 24 and for establishing a communication channel 34. Will be understood. It will further be appreciated that the input device and output device need not be included in the same user device 38. For example, text for the user to speak can be displayed on a television that communicates with the system 24 over the network, while voice is recorded by the user using a smartphone that communicates with the system 24 over the network.

  In any case, as described above, the input module 28 is configured to receive text in a specified language or other language from the user and to receive the voice of the user speaking the text in the specified language. The processing module 30 processes the text and utters the text to derive expected characteristics (eg, fluency) of the utterance of the received text from the input module 28 using, for example, a speech recognition algorithm. It is configured to process the received speech from the input module 28 in order to derive the user's characteristics. The processing module 30 is also used to parse the text into one or more sentences. That is, the received text containing a collection of words is parsed into each sentence by the processing module 30 and the text is searched in turn for the appearance of ending quoted sentence punctuation marks such as terminators, question marks, and exclamation marks. Thus, in use, the user is prompted to speak the first sentence of each parsed sentence, and the voice of the user speaking the first sentence of each sentence of the text is input by the input module 28. Receive, and so on.

  In one example, the sentence: He Said, “I am not done!” The he continued speaking. Is a single sentence: He said, “I am not done”, the he continued speaking. Parse into But the next sentence: He said, I am not done! The he continued speaking. Are two sentences: He Said, I am not done! 2. The he continued speaking. And parse them to prompt the user to speak these sentences. Furthermore, once the aggregate of words from the input text is parsed into individual sentences, each sentence is converted into a speech recognition grammar that matches the expected characteristics of the sentence. Sentences are presented visually or acoustically to the user, prompting the user to speak what they have read or heard, and the user's voice response is sent from the input module 28 along with the sentence grammar for analysis. It is sent to the speech recognition device executed by the processing module 30.

  Accordingly, the processing module 30 determines the accuracy of the user speaking the text with respect to the characteristics, such as the expected characteristics of the text utterance imposed by the speech recognition vendor, for example, the characteristics of the user speaking the text. Compare with The output module 32 receives the accuracy determination result from the processing module 30 and outputs feedback to the user indicating the accuracy of the user speaking the text.

  FIG. 3 shows a system 36 for assisting in improving user utterances in a specified language, including a user device 38 for inputting and outputting information to the processor 26 via a network 40 such as the Internet. As mentioned above, in other embodiments not shown, it is contemplated that the system 24 is self-contained and includes suitable components for inputting and outputting information to the processor 26 for performing the method 10. Nevertheless, the processor 26 is illustrated as executing by a server 54 accessible via the Internet 40. It is contemplated that the processor 26 and memory 52 may comprise, for example, a virtual server or may be implemented as a cloud service that spans many physical servers. In one example, the server 54 is accessible via a uniform resource locator (URL) that a user can embed in a website for access via the user device 38. In another example, the network 40 is a local area network (LAN), and the user device 38 communicates with a server, for example via WiFi.

  The user device 38 shown in FIG. 3 includes a display 44 configured to display text in a specified language or other language to the user. In the example illustrated in FIG. 3, the displayed text is “Hey Joe, what are you doing”, and it can be seen that the user device is a tablet computer having a touch screen or a gesture reading function. As described above, other user devices such as smartphones and personal computers may be used with the system 36.

  With reference to an example using this embodiment, the user uses the input means 46 in the form of a touch screen keyboard as part of the text for the user to speak to assist the user in learning English. Text: “Hey Joe, what are you doing” has been entered. Once the user has finished typing text or pasting text from another document, the user can press the “Submit” button (not shown) and send the text to the server 54 via the Internet 40. Become. The entered text is transmitted to the input module 28 over communication channels 42 and 34 over the Internet 40 so that the processing module 30 can process the text to derive the expected characteristics of the text utterance. As described above, the processing module 30 may send the text to the user to speak one sentence at a time in order to receive feedback indicating how accurately the text was spoken at the conclusion of each sentence. Analyze each sentence for display. Further, in one embodiment, the processing module 30 converts each sentence into a context free grammar, which is the target speech recognition device employed by the processing module 30 to analyze the utterance of the user who is speaking the sentence. This is the required syntax. That is, in this embodiment, the grammar shows the structure of the text put together for transport to the speech recognition device.

  Referring to the same example, the text “Hey Joe, what are you doing” is output back to the user device 38 via the output module 32 and displayed on the display 44 as a reminder signal for speaking the first sentence. That is, the reminder signal includes the text of the first sentence for speaking. The microphone 48 is used to record the user who is speaking this sentence, and the voice signal of the user who is speaking the sentence is transmitted to the input module 28 via the Internet 40. This allows the processing module 30 to process the audio signal to derive the characteristics of the user speaking the text. The processing module 30 compares the derived characteristics of the speech signal with the expected characteristics to determine the accuracy of the user speaking the text using the target speech recognition algorithm.

  The output module 32 provides feedback to the user via the Internet 40 as a color for displaying feedback indicating the accuracy determined by the processing module 30 on the display 44 indicating various levels of confidence values for the determined accuracy. Output. As described above, green is used to indicate a confidence value associated with an accuracy that is “very accurate (eg, greater than 80% match)”. In another example, the output module 32 outputs feedback indicating the accuracy determined by the processing module 30 to the user as a voice output from the speaker 50 of the user device 38. For example, the voice feedback states “very accurate” through the speaker 50 when the processing module 30 determines a confidence value higher than 80%. As described above, the display 44 can also be configured to display other indicators that indicate various levels of confidence values for the determined accuracy, such as numerical and letter grades. Further, in one embodiment, the processing module 30 includes a speech synthesis algorithm as described above, where the output module 32 outputs to the user from the speaker 50 to further assist the user in improving speech. It can be configured to output an expected utterance composed of text to synthesize.

  After outputting the feedback, the output module 32 outputs the next sentence that the processing module 30 has parsed (not shown) from the text entered by the user as a further prompting signal for the user to speak the second sentence. Further configuration to A reminder signal containing the text of the second sentence is also displayed on the display 44 and the microphone 48 is used to record the user speaking the sentence. As in the first sentence, the voice signal of the user speaking the sentence is input module so that the processing module 30 can process the voice signal to derive the characteristics of the user speaking the text. 28. The processing module 30 compares the derived characteristics of the speech signal with the expected characteristics to determine the accuracy of the user speaking the text, and the output module 32 indicates the accuracy determined by the processing module 30. Feedback is output to the user as a color to be displayed on the display 44. The output module outputs the next sentence, and repeats the process until all sentences analyzed from the text input by the user are uttered or until the user finishes the process at an earlier time.

  As described above, the memory 52 can be used to record a voice signal of a user speaking a text in a specified language. Accordingly, the processing module 30 can further process each example of the audio signal recorded in the memory 52 to derive and refine one or more expected characteristics of the text utterance. This is because the expected pronunciation of the text can be determined from an analysis of the recording of the user speaking the text. In addition, the recorded utterance is accompanied by data indicating the particular user in the memory 52 so that the recording can be retrieved for later analysis, eg, to show the progress of the particular user in the English utterance. Save to. Those skilled in the art will appreciate that the method and system can be employed for users who wish to improve speech in other languages such as French, Chinese, Japanese.

  Further aspects of the method will become apparent from the above description of the system. One skilled in the art will also appreciate that the method can be embodied in program code. This program code can be provided in several ways, for example on a tangible computer readable medium such as a disk or memory, or as a data signal or data file.

  Those skilled in the art will recognize that many changes, additions, and / or modifications may be made without departing from the spirit and scope of the invention.

  Where any prior art is referred to herein, that reference is not an admission that the prior art forms part of the general general knowledge in that technical field in any country. Will be understood.

  The present invention may be used as a basis for priority in respect of one or more future applications, and the claims of such future applications are subject to any one feature or feature described in this application. Combinations may also be targeted. As such, future applications include one or more of the following claims, which are given in an example manner in any future application, and are not limited as to what is claimed.

Claims (35)

A method of helping to improve user utterances in a specified language,
Receiving from the user text in the specified language or other language for the user to speak in the specified language;
Processing the text to derive one or more expected characteristics of the text utterance in the specified language;
Receiving an audio signal of the user speaking the text in the specified language;
Processing the speech signal to derive one or more characteristics of the user speaking the text;
In order to determine the accuracy of the user speaking the text, the one or more characteristics of the user speaking the text are changed to the one or more expected characteristics of the speech of the text. Comparing with the corresponding characteristic of
Outputting to the user feedback indicating the accuracy of the user speaking the text;
Parsing the text into one or more sentences or phrases of words for the user to speak;
Receiving an audio signal of the user speaking the sentence or clause of the text;
Outputting the feedback to the user indicating the accuracy of the user speaking the sentence or clause of the text.   Outputting a prompting signal for causing the user to speak the first one of the sentence or clause; and receiving a voice signal of the user speaking the first one of the sentence or clause of the text The method of claim 2, further comprising:   The method of claim 2, further comprising outputting the feedback to the user indicating the accuracy of the user speaking the first one of the sentence or clause of the text.   After outputting the feedback to the user indicating the accuracy of the user speaking the first one of the sentence or clause of the text, a second one of the sentence or clause is given to the user. 4. The method of claim 3, further comprising the step of outputting a further prompting signal for the user to speak.   The method of claim 2, wherein the reminder signal includes text of the first of the sentences or clauses for display on a display of a user device to the user.   The method of claim 1, further comprising generating a speech recognition grammar of the text.   7. A method according to any preceding claim, further comprising the step of processing the text to derive a synthesized expected utterance of the specified language text.   8. The method of claim 7, further comprising outputting the synthesized expected utterance of the text to the user via a user device speaker.   The accuracy includes a confidence value associated with the accuracy of the one or more characteristics of the user speaking the text corresponding to the one or more expected characteristics of the text utterance. Item 9. The method according to any one of Items 1 to 8.   The method of claim 9, wherein the feedback includes colors that indicate various confidence values of the accuracy.   11. A method according to any preceding claim, further comprising the step of processing the text to derive the predetermined prohibited word in the text such that a predetermined prohibited word can be censored.   The one or more expected characteristics of the utterance of the text and the one or more characteristics of the user speaking the text include at least one of pronunciation, fluency, and prosodic features; The method according to claim 1.   13. The method of any preceding claim, further comprising processing the speech signal to derive the one or more characteristics of the user speaking the text using a speech recognition algorithm. Method.   14. A method as claimed in any preceding claim, further comprising processing the text to derive the one or more expected properties of the text using a speech synthesis algorithm. A system for supporting improvement of user utterances in a specified language,
A processor,
Receiving from the user text in the specified language or other language for the user to speak in the specified language; and
An input module configured to receive a speech moderation of the user speaking the text in the specified language;
Processing the text to derive one or more expected characteristics of the utterance of the text in the specified language;
Processing the speech signal to derive one or more characteristics of the user speaking the text; and
In order to determine the accuracy of the user speaking the text, the one or more characteristics of the user speaking the text are changed to the one or more expected characteristics of the speech of the text. A processing module configured to be compared with a corresponding characteristic of
An output module configured to output to the user feedback indicating the accuracy of the user speaking the text;
The processing module is further configured to parse the text into one or more sentences or clauses of a word;
The input module is further configured to receive an audio signal of the user speaking the one or more sentences or clauses of the text; and
The output module is further configured to output the feedback to the user indicating the accuracy of the user speaking the sentence or passage of the text.   The system of claim 15, wherein the output module is further configured to output a reminder signal to cause a user to speak the first of the sentences or phrases.   The system of claim 16, wherein the input module is further configured to receive an audio signal of the user speaking the first one of the sentence or clause of the text.   The output module is further configured to output the feedback to the user indicating the accuracy of the user speaking the first one of the sentence or clause of the text. 17. The system according to 17.   The output module outputs the feedback indicating the accuracy of the user speaking the first one of the sentence or clause of the text to a second of the sentence or clause after outputting to the user. The system of claim 18, further configured to output a further reminder signal to cause the user to speak.   The system of claim 16, wherein the reminder signal includes text of the first of the sentences or clauses for display on a display of a user device to the user.   The system of claim 15, wherein the processing module is further configured to generate a speech recognition grammar for the text.   22. A system according to any of claims 15 to 21, wherein the processing module is further configured to process the text to derive a synthesized expected utterance of the text in the specified language.   23. The system of claim 22, wherein the output module is further configured to output the synthesized expected utterance of the text to the user via a user device speaker.   The accuracy includes a confidence value associated with the accuracy of the one or more characteristics of the user speaking the text corresponding to the one or more expected characteristics of the text utterance. Item 24. The system according to any one of Items 15 to 23.   25. The system of claim 24, wherein the feedback includes colors that indicate various confidence values of the accuracy.   26. Any of claims 15-25, wherein the processing module is further configured to process the text to derive the predetermined prohibited word in the text such that the predetermined prohibited word can be censored. The system described in Crab.   27. A system according to any of claims 15 to 26, wherein the system further comprises a memory for recording the audio signal of the user speaking the text.   28. The processing module of claim 27, wherein the processing module is further configured to process an example of the audio signal recorded in the memory to derive the one or more expected characteristics of the text utterance. system.   The one or more expected characteristics of the utterance of the text and the one or more characteristics of the user speaking the text include at least one of pronunciation, fluency, and prosodic features; A system according to any of claims 15 to 28. A system for supporting improvement of user utterances in a specified language,
A display configured to display text in the specified language or other languages;
Text input means configured to input the text of the specified language or the other language by the user for the user to speak in the specified language;
A microphone configured to input an audio signal of the user speaking the text in the specified language;
A processor,
Receiving the text entered in the specified language or the other language; and
An input module configured to receive the audio signal of the user speaking the text in the specified language;
Processing the text to derive one or more expected characteristics of the utterance of the text in the specified language;
Processing the speech signal to derive one or more characteristics of the user speaking the text; and
In order to determine the accuracy of the user speaking the text, the one or more characteristics of the user speaking the text are changed to the one or more expected characteristics of the speech of the text. A processing module configured to be compared with a corresponding one of
An output module configured to output to the user feedback indicating the accuracy of the user speaking the text;
With
The processing module is further configured to parse the text into one or more sentences or clauses of a word;
The input module is further configured to receive an audio signal of the user speaking the one or more sentences or clauses of the text; and
The output module is further configured to output the feedback to the user indicating the accuracy of the user speaking the sentence or passage of the text.   32. The system of claim 30, wherein the display is further configured to display the feedback to the user.   32. The system of claim 31, further comprising a server including the processor in data communication over a network with a user device including the display, the text input means, and the microphone.   Computer program code that, when executed, executes the method according to any of claims 1-14.   34. A tangible computer readable medium containing the program code of claim 33.   34. A data file containing the program code of claim 33.

Images